超宽带钢典型板形缺陷向量提取方法

Extracting method of typical shape defect vector for super-wide strip steel

  • 摘要: 提出了从带钢板形数据中提取典型板形缺陷向量的方法,即先对带钢板形缺陷进行初级分类,再根据带钢应力之间的相似度提取出单位典型板形向量,最后采用k均值聚类算法求解出典型比值系数.针对提取出的带钢单位典型板形向量而设计出的单侧倒角工作辊,经过大规模的轧制试验,所轧宽带钢板形质量较好,综合平坦度均小于5 IU(1 IU=10-5),基本消除边中复合浪为主的复杂浪形缺陷.

     

    Abstract: This article introduces a method of extracting the typical defect vector of strip shape. In the method, firstly the strip shape defects are primarily classified, then the unit typical strip shape vector is extracted according to similarity among strip stresses, and finally the typical ratio coefficient is solved with the k-means cluster approach. Aiming at the unit typical strip shape vector extracted, a taper work roll was designed and put into a large-scale trial. It is shown that the shape quality of rolled strips is better than before. The comprehensive flatness is below 5 IU, and side and middle composite buckles are basically eliminated.

     

/

返回文章
返回